APP下载

基于大数据分析的电能损耗预测模型的仿真分析

2016-05-14张明明

现代电子技术 2016年6期

张明明

摘 要: 对电能损耗进行预测的过程中,传统的预测方法,由于影响电能损耗的因素之间具有非常复杂的非线性与强相关性,导致很难得到高精度的预测结果。提出一种基于大数据分析的电能损耗预测模型,建立电能损耗预测单元模型,分析发电器模型、变电器模型和输电器模型建立的过程,引入灰色综合关联度对大数据进行分析,对各个组成区域序列累加得到整体区域电能损耗情况,并进行微分处理,通过GM(1,1)的转换实现对整体电能损耗预测模型的构建。仿真实验结果表明,所提模型具有较高的预测精度,能够很好地避免非线性因素对模型的干扰,为电力企业制定合理的电能损耗计划提供可靠依据。

关键词: 大数据分析; 电能损耗预测; 灰度关联; 仿真实验

中图分类号: TN911?34;TM711 文献标识码: A 文章编号: 1004?373X(2016)06?0153?04

Simulation analysis of electric energy loss prediction model based on big data analysis

ZHANG Mingming

(Institute of Smart Grid, CSG Science Academy, Guangzhou 510080, China)

Abstract: The traditional prediction method in the process of electric energy loss prediction is hard to obtain the high?precision prediction results due to the very complicated nonlinear and strong correlation between factors influencing on the electric energy loss. An electric energy loss prediction model based on big data analysis is put forward. The unit model of the electric energy loss prediction was established, and the building process of generator model, electric transducer model and electric transmitter model was analyzed. The gray synthesis correlation degree was introduced to analyze the big data. The electric energy loss situation of the whole region was obtained by sequence accumulation of each sub?region, and then a differential treatment was conducted. The construction of the whole electric energy loss prediction model was realized by GM (1,1) transformation. The simulation results show that the proposed model has high prediction accuracy, and can better avoid the interference from nonlinear factor. The model can provide a reliable basis for the power enterprises to make the reasonable electric energy loss plan.

Keywords: big data analysis; electric energy loss prediction; gray?scale correlation; simulation experiment

0 引 言

近年来,随着人们物质条件的逐渐发展,能源越来越紧缺,减少电能损耗成为亟需解决的问题[1?2]。随着电网运行管理的智能化,电能损耗计算对算法的精确程度要求越来越高,如何有效准确地进行电能损耗预测,对电网企业开展降损节能工作具有重要意义,其已经成为相关学者研究的重点课题[3?5]。目前,针对电能损耗的预测方法主要包括潮流方法、神经网络方法和支持向量机方法等,相关研究也取得了一定的成果。文献[6]提出一种基于BP神经网络的电能损耗预测模型,通过神经网络对电能损耗进行预测,但该方法存在局部收敛、收敛效率低以及隐含层神经元选择复杂等弊端,而且该预测模型还需要大样本容量。……

登录APP查看全文